On Tue, Jun 19, 2012 at 2:26 PM, Tejun Heo [off-list ref] wrote:

On Tue, Jun 19, 2012 at 02:20:59PM -0700, Tejun Heo wrote:

quoted

Something like the following should fix it.

diff --git a/mm/memblock.c b/mm/memblock.c
index 32a0a5e..2770970 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c

@@ -148,11 +148,15 @@ phys_addr_t __init_memblock memblock_find_in_range(phys_addr_t start,

  */
 int __init_memblock memblock_free_reserved_regions(void)
 {
+#ifndef CONFIG_DEBUG_PAGEALLOC
      if (memblock.reserved.regions == memblock_reserved_init_regions)
              return 0;

      return memblock_free(__pa(memblock.reserved.regions),
               sizeof(struct memblock_region) * memblock.reserved.max);
+#else
+     return 0;
+#endif

BTW, this is just ugly and I don't think we're saving any noticeable
amount by doing this "free - give it to page allocator - reserve
again" dancing.  We should just allocate regions aligned to page
boundaries and free them later when memblock is no longer in use.

if it is that case, that change could fix other problem problem too.

--- during the one free reserved.regions could double the array.

please check attached patch.

Yinghai